Brier
Average over finished matches: how close the model’s win/draw/loss percentages were to what actually happened. 0 = perfect. Lower is better. Does not add ranking points — only breaks ties when two models have the same total.
0.539
Log loss
Average over finished matches: did the model assign enough probability to the outcome that really happened? Being very confident but wrong hurts a lot. Lower is better. Does not add ranking points.
Brier
Average over finished matches: how close the model’s win/draw/loss percentages were to what actually happened. 0 = perfect. Lower is better. Does not add ranking points — only breaks ties when two models have the same total.
0.538
Log loss
Average over finished matches: did the model assign enough probability to the outcome that really happened? Being very confident but wrong hurts a lot. Lower is better. Does not add ranking points.
Brier
Average over finished matches: how close the model’s win/draw/loss percentages were to what actually happened. 0 = perfect. Lower is better. Does not add ranking points — only breaks ties when two models have the same total.
0.550
Log loss
Average over finished matches: did the model assign enough probability to the outcome that really happened? Being very confident but wrong hurts a lot. Lower is better. Does not add ranking points.
Brier
Average over finished matches: how close the model’s win/draw/loss percentages were to what actually happened. 0 = perfect. Lower is better. Does not add ranking points — only breaks ties when two models have the same total.
0.525
Log loss
Average over finished matches: did the model assign enough probability to the outcome that really happened? Being very confident but wrong hurts a lot. Lower is better. Does not add ranking points.
Brier
Average over finished matches: how close the model’s win/draw/loss percentages were to what actually happened. 0 = perfect. Lower is better. Does not add ranking points — only breaks ties when two models have the same total.
0.526
Log loss
Average over finished matches: did the model assign enough probability to the outcome that really happened? Being very confident but wrong hurts a lot. Lower is better. Does not add ranking points.
Brier
Average over finished matches: how close the model’s win/draw/loss percentages were to what actually happened. 0 = perfect. Lower is better. Does not add ranking points — only breaks ties when two models have the same total.
0.539
Log loss
Average over finished matches: did the model assign enough probability to the outcome that really happened? Being very confident but wrong hurts a lot. Lower is better. Does not add ranking points.
Brier
Average over finished matches: how close the model’s win/draw/loss percentages were to what actually happened. 0 = perfect. Lower is better. Does not add ranking points — only breaks ties when two models have the same total.
0.539
Log loss
Average over finished matches: did the model assign enough probability to the outcome that really happened? Being very confident but wrong hurts a lot. Lower is better. Does not add ranking points.
Brier
Average over finished matches: how close the model’s win/draw/loss percentages were to what actually happened. 0 = perfect. Lower is better. Does not add ranking points — only breaks ties when two models have the same total.
0.530
Log loss
Average over finished matches: did the model assign enough probability to the outcome that really happened? Being very confident but wrong hurts a lot. Lower is better. Does not add ranking points.
Brier
Average over finished matches: how close the model’s win/draw/loss percentages were to what actually happened. 0 = perfect. Lower is better. Does not add ranking points — only breaks ties when two models have the same total.
0.542
Log loss
Average over finished matches: did the model assign enough probability to the outcome that really happened? Being very confident but wrong hurts a lot. Lower is better. Does not add ranking points.
Brier
Average over finished matches: how close the model’s win/draw/loss percentages were to what actually happened. 0 = perfect. Lower is better. Does not add ranking points — only breaks ties when two models have the same total.
0.571
Log loss
Average over finished matches: did the model assign enough probability to the outcome that really happened? Being very confident but wrong hurts a lot. Lower is better. Does not add ranking points.
Brier
Average over finished matches: how close the model’s win/draw/loss percentages were to what actually happened. 0 = perfect. Lower is better. Does not add ranking points — only breaks ties when two models have the same total.
0.530
Log loss
Average over finished matches: did the model assign enough probability to the outcome that really happened? Being very confident but wrong hurts a lot. Lower is better. Does not add ranking points.
Brier
Average over finished matches: how close the model’s win/draw/loss percentages were to what actually happened. 0 = perfect. Lower is better. Does not add ranking points — only breaks ties when two models have the same total.
0.548
Log loss
Average over finished matches: did the model assign enough probability to the outcome that really happened? Being very confident but wrong hurts a lot. Lower is better. Does not add ranking points.
Brier
Average over finished matches: how close the model’s win/draw/loss percentages were to what actually happened. 0 = perfect. Lower is better. Does not add ranking points — only breaks ties when two models have the same total.
0.553
Log loss
Average over finished matches: did the model assign enough probability to the outcome that really happened? Being very confident but wrong hurts a lot. Lower is better. Does not add ranking points.
Brier
Average over finished matches: how close the model’s win/draw/loss percentages were to what actually happened. 0 = perfect. Lower is better. Does not add ranking points — only breaks ties when two models have the same total.
0.555
Log loss
Average over finished matches: did the model assign enough probability to the outcome that really happened? Being very confident but wrong hurts a lot. Lower is better. Does not add ranking points.
Brier
Average over finished matches: how close the model’s win/draw/loss percentages were to what actually happened. 0 = perfect. Lower is better. Does not add ranking points — only breaks ties when two models have the same total.
0.550
Log loss
Average over finished matches: did the model assign enough probability to the outcome that really happened? Being very confident but wrong hurts a lot. Lower is better. Does not add ranking points.
Brier
Average over finished matches: how close the model’s win/draw/loss percentages were to what actually happened. 0 = perfect. Lower is better. Does not add ranking points — only breaks ties when two models have the same total.
0.587
Log loss
Average over finished matches: did the model assign enough probability to the outcome that really happened? Being very confident but wrong hurts a lot. Lower is better. Does not add ranking points.